Passive Muscle Force?

Hello,

I have a muscle modeled using AnyMuscleModel3E, it is connect through
viapointmuscle and the distance between the two points it connects is
0.5m. I gave the following properties for it:

AnyMuscleModel3E MuscleModelC = {
F0 = 100;
Lfbar = 0.5;
Lt0 = 0.5;
Epsilonbar = 0.05;
Gammabar = 30*pi/180; //Pennation Angle
Fcfast = 0.4; //Fraction of fast twitch fibers
Jt = 3.0; //Elasiticity of tendon (serial)
Jpe = 3.0; //Elasiticity of tendon (Parallel)
K1 = 2;
K2 = 8;
PEFactor = 5;
};

It contract to lift up a segment connected to a joint-no external load.

My problem is that the passive force appear throughout the
simulation—something not desired.

I calibrated the muscle at the initial position, and try to change the
Lfbar and Lt0 value, but the passive force still exist throughout the
simulation.

I would like to have a passive force to act in the following way: at
the initial stage, there is no passive force, as the muscle contract,
the passive force store up(like compressing a spring) and act in a
direction opposing the contracting force(?), and as the motion change
direction, the passive force release(like let go a compressed spring).

Could this be done? What parameter(s) should I change/adjust? Or my
concept about the muscle modeling is wrong? Detail explaination
greately appreciated.

Thanks,
Leng-Feng

Hi Leng-Feng

I have uploaded a small example named “MuscleModel3eLengFeng.any” to
the file section of the group.

It is a small model to demonstrate the 3 element hill muscle model.
The model consist of one mass which is attached to the global ref
using a prismatic joint the mass is hanging in a muscle which will
lift the mass upward when running the anlysis.

To run the model please calibrate first then run the inverse anal.

In the current state of the model the mass of the segment is 6 kg
This will give activity slightly above one
The passive force will have a value in the two first steps but it is
very small
time step1 Fp=0.0155 N
time step2 Fp=0.0036 N
time step3 Fp=0

If you now try to change the mass of the segment to almost zero
Mass=0.05 kg
Then rerun the analysis you will get these results for the first
steps
Activity = 0.053
Passive Force Fp=0.131
The passive force will continue to have values until timestep 11

It may seem strange that the passive force is biggest in the case
with the small mass, here is the explanation:

When the calibration is preformed we optimize the muscle to have its
optimal strength at the length in which it is calibrated. We simply
take the overall length from origin to insertation and subtract the
lenght of the contractile element to obtain the tendon length.
However we also take into account the strain of the muscle when it
is maximally loaded, when we do the calibration.
This means that teoretically the passive force should be zero when
the activity is one, this is also what the results of the analysis
shows. The explanation of why it do not give exactly zero is the
non linearities in the muscle model.

Then the question why do the model with the small mass give a larger
passive force?
When the muscle is loaded with the weight of a small mass, the
tendon will be shorter than the calibrated length because in the
calibrated length the strain of the tendon for activity one
is taken into account. In other words: the muscle is calibrated for
activity one. Since we in both cases drive the overall lengh of the
muscle to the same length the passive element will be more strecthed
in the case where we have the small mass, because the tendon is not
loaded so much.

So you may ask why did we choose to take into account the tendon
strain when we calibrate the muscle?
This we did because we wanted to have the optimum strength exactly
at the calibrated length this would not have been the result
otherwise.

The passive element should work exactly as you describe in the end
of your email, but when the muscle is not loaded very much the
passive force will not be zero at the calibrated length. Furthermore
the parameters of your model enhance this phenomean because your
tendon length is extremely long.

I hope this answers your question otherwise please do not hesitate
to write again.

Best regards
AnyBody Support

— In anyscript@yahoogroups.com, “Leng-Feng Lee” <lengfenglee@…>
wrote:
>
> Hello,
>
> I have a muscle modeled using AnyMuscleModel3E, it is connect
through
> viapointmuscle and the distance between the two points it connects
is
> 0.5m. I gave the following properties for it:
>
> AnyMuscleModel3E MuscleModelC = {
> F0 = 100;
> Lfbar = 0.5;
> Lt0 = 0.5;
> Epsilonbar = 0.05;
> Gammabar = 30*pi/180; //Pennation Angle
> Fcfast = 0.4; //Fraction of fast twitch fibers
> Jt = 3.0; //Elasiticity of tendon (serial)
> Jpe = 3.0; //Elasiticity of tendon (Parallel)
> K1 = 2;
> K2 = 8;
> PEFactor = 5;
> };
>
> It contract to lift up a segment connected to a joint-no external
load.
>
> My problem is that the passive force appear throughout the
> simulation—something not desired.
>
> I calibrated the muscle at the initial position, and try to change
the
> Lfbar and Lt0 value, but the passive force still exist throughout
the
> simulation.
>
> I would like to have a passive force to act in the following way:
at
> the initial stage, there is no passive force, as the muscle
contract,
> the passive force store up(like compressing a spring) and act in a
> direction opposing the contracting force(?), and as the motion
change
> direction, the passive force release(like let go a compressed
spring).
>
> Could this be done? What parameter(s) should I change/adjust? Or my
> concept about the muscle modeling is wrong? Detail explaination
> greately appreciated.
>
> Thanks,
> Leng-Feng
>

Hi,

Thanks for the quick reply and the demo.

To better understand my problem at hand, I have uploaded a model call
“OneSegModel.any”, which uses two muscles to move a pivoted segment
swinging back and forth.

The muscles are calibrated at its initial position, then run the
simulation. After the simulation, the muscle forces Ft/Fm both have
the similar profile, as the Passive force Fp.

It is strange that when the segment swing to the right, this movement
should be cause by the contraction of the right muscle, but instead
the muscle force profile of the right muscle is zero at this
interval–but when look at muscle force profile of the left muscle,
it has a bell curve. I wonder why this is so? It act as if the ‘swing
to the right’ motion is cause (push) by the passive force in the Left
muscle (?).

I guess (?) my muscle model parameters didn’t set up correctly such
that after the calibration, there are still a lot of passive force in
the muscle? Or what could be the problem?

And, am I right that the “negative valued” force in the muscle will
shown as ‘zero’ in the plotted force profile? If so, does the anybody
computation machines take the ‘negative valued’ force into account
when it does the subsequence computation or it will take ‘zero’ as a
value?

Many thanks,
Leng-Feng

— AnyBody Support <support@anybodytech.com> wrote:

> Hi Leng-Feng
>
> I have uploaded a small example named “MuscleModel3eLengFeng.any”
> to
> the file section of the group.
>
> It is a small model to demonstrate the 3 element hill muscle model.
> The model consist of one mass which is attached to the global ref
> using a prismatic joint the mass is hanging in a muscle which will
> lift the mass upward when running the anlysis.
>
> To run the model please calibrate first then run the inverse anal.
>
> In the current state of the model the mass of the segment is 6 kg
> This will give activity slightly above one
> The passive force will have a value in the two first steps but it
> is
> very small
> time step1 Fp=0.0155 N
> time step2 Fp=0.0036 N
> time step3 Fp=0
>
> If you now try to change the mass of the segment to almost zero
> Mass=0.05 kg
> Then rerun the analysis you will get these results for the first
> steps
> Activity = 0.053
> Passive Force Fp=0.131
> The passive force will continue to have values until timestep 11
>
> It may seem strange that the passive force is biggest in the case
> with the small mass, here is the explanation:
>
> When the calibration is preformed we optimize the muscle to have
> its
> optimal strength at the length in which it is calibrated. We simply
>
> take the overall length from origin to insertation and subtract the
>
> lenght of the contractile element to obtain the tendon length.
> However we also take into account the strain of the muscle when it
> is maximally loaded, when we do the calibration.
> This means that teoretically the passive force should be zero when
> the activity is one, this is also what the results of the analysis
> shows. The explanation of why it do not give exactly zero is the
> non linearities in the muscle model.
>
> Then the question why do the model with the small mass give a
> larger
> passive force?
> When the muscle is loaded with the weight of a small mass, the
> tendon will be shorter than the calibrated length because in the
> calibrated length the strain of the tendon for activity one
> is taken into account. In other words: the muscle is calibrated for
>
> activity one. Since we in both cases drive the overall lengh of the
>
> muscle to the same length the passive element will be more
> strecthed
> in the case where we have the small mass, because the tendon is not
>
> loaded so much.
>
> So you may ask why did we choose to take into account the tendon
> strain when we calibrate the muscle?
> This we did because we wanted to have the optimum strength exactly
> at the calibrated length this would not have been the result
> otherwise.
>
> The passive element should work exactly as you describe in the end
> of your email, but when the muscle is not loaded very much the
> passive force will not be zero at the calibrated length.
> Furthermore
> the parameters of your model enhance this phenomean because your
> tendon length is extremely long.
>
>
> I hope this answers your question otherwise please do not hesitate
> to write again.
>
> Best regards
> AnyBody Support
>
>
>
>
>
>
> — In anyscript@yahoogroups.com, “Leng-Feng Lee” <lengfenglee@…>
>
> wrote:
> >
> > Hello,
> >
> > I have a muscle modeled using AnyMuscleModel3E, it is connect
> through
> > viapointmuscle and the distance between the two points it
> connects
> is
> > 0.5m. I gave the following properties for it:
> >
> > AnyMuscleModel3E MuscleModelC = {
> > F0 = 100;
> > Lfbar = 0.5;
> > Lt0 = 0.5;
> > Epsilonbar = 0.05;
> > Gammabar = 30*pi/180; //Pennation Angle
> > Fcfast = 0.4; //Fraction of fast twitch fibers
> > Jt = 3.0; //Elasiticity of tendon (serial)
> > Jpe = 3.0; //Elasiticity of tendon (Parallel)
> > K1 = 2;
> > K2 = 8;
> > PEFactor = 5;
> > };
> >
> > It contract to lift up a segment connected to a joint-no external
>
> load.
> >
> > My problem is that the passive force appear throughout the
> > simulation—something not desired.
> >
> > I calibrated the muscle at the initial position, and try to
> change
> the
> > Lfbar and Lt0 value, but the passive force still exist throughout
>
> the
> > simulation.
> >
> > I would like to have a passive force to act in the following way:
>
> at
> > the initial stage, there is no passive force, as the muscle
> contract,
> > the passive force store up(like compressing a spring) and act in
> a
> > direction opposing the contracting force(?), and as the motion
> change
> > direction, the passive force release(like let go a compressed
> spring).
> >
> > Could this be done? What parameter(s) should I change/adjust? Or
> my
> > concept about the muscle modeling is wrong? Detail explaination
> > greately appreciated.
> >
> > Thanks,
> > Leng-Feng
> >
>
>
>
>
>


Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com

Hi Leng-Feng

Nice little model i have taken a look at it.

There is nothing worng with the muscle parameters, it is the
dynamics effects of teh kinematics that makes the trick here.

If you change the driver sligthly to be dependent on the tEnd value
of the analysis like this

AnyKinEqFourierDriver Driver = {
Type = Sin;
Freq = 2/Main.OneLegStudy.tEnd ; //# of cycles complete
in 1 sec
A = {{0,1/2}}; //Control the amplitude of swing
B = {{0,0}};
AnyRevoluteJoint &Joint = Main.OneLegModel.Jnts.LegJoint;
Reaction.Type = {Off};
};

If gradually increase the tEnd time of the analysis (this is a value
you can set in your study) you will see that about tEnd=2.5 sec the
right muscles statrts to be activated first.When you increase tEnd
to 10 sec you have the bell shapes you are looking for i belive.

So when the motion is too fast the rigth muscle will be actiated
first because it has to brake the motion and change the direction of
it.

Our normal muscles can not have pushing forces this not possible in
any way, but the AnyGeneralMuscles can be set to act like a pushing
element, this is done be setting their ForceDirection paramter to 1.

Best regards
AnyBody Support

— In anyscript@yahoogroups.com, Leng-Feng Lee <lengfenglee@…>
wrote:
>
> Hi,
>
> Thanks for the quick reply and the demo.
>
> To better understand my problem at hand, I have uploaded a model
call
> “OneSegModel.any”, which uses two muscles to move a pivoted segment
> swinging back and forth.
>
> The muscles are calibrated at its initial position, then run the
> simulation. After the simulation, the muscle forces Ft/Fm both have
> the similar profile, as the Passive force Fp.
>
> It is strange that when the segment swing to the right, this
movement
> should be cause by the contraction of the right muscle, but instead
> the muscle force profile of the right muscle is zero at this
> interval–but when look at muscle force profile of the left muscle,
> it has a bell curve. I wonder why this is so? It act as if
the ‘swing
> to the right’ motion is cause (push) by the passive force in the
Left
> muscle (?).
>
> I guess (?) my muscle model parameters didn’t set up correctly such
> that after the calibration, there are still a lot of passive force
in
> the muscle? Or what could be the problem?
>
> And, am I right that the “negative valued” force in the muscle will
> shown as ‘zero’ in the plotted force profile? If so, does the
anybody
> computation machines take the ‘negative valued’ force into account
> when it does the subsequence computation or it will take ‘zero’ as
a
> value?
>
> Many thanks,
> Leng-Feng
>
> — AnyBody Support <support@…> wrote:
>
> > Hi Leng-Feng
> >
> > I have uploaded a small example named “MuscleModel3eLengFeng.any”
> > to
> > the file section of the group.
> >
> > It is a small model to demonstrate the 3 element hill muscle
model.
> > The model consist of one mass which is attached to the global
ref
> > using a prismatic joint the mass is hanging in a muscle which
will
> > lift the mass upward when running the anlysis.
> >
> > To run the model please calibrate first then run the inverse
anal.
> >
> > In the current state of the model the mass of the segment is 6 kg
> > This will give activity slightly above one
> > The passive force will have a value in the two first steps but it
> > is
> > very small
> > time step1 Fp=0.0155 N
> > time step2 Fp=0.0036 N
> > time step3 Fp=0
> >
> > If you now try to change the mass of the segment to almost zero
> > Mass=0.05 kg
> > Then rerun the analysis you will get these results for the first
> > steps
> > Activity = 0.053
> > Passive Force Fp=0.131
> > The passive force will continue to have values until timestep 11
> >
> > It may seem strange that the passive force is biggest in the
case
> > with the small mass, here is the explanation:
> >
> > When the calibration is preformed we optimize the muscle to have
> > its
> > optimal strength at the length in which it is calibrated. We
simply
> >
> > take the overall length from origin to insertation and subtract
the
> >
> > lenght of the contractile element to obtain the tendon length.
> > However we also take into account the strain of the muscle when
it
> > is maximally loaded, when we do the calibration.
> > This means that teoretically the passive force should be zero
when
> > the activity is one, this is also what the results of the
analysis
> > shows. The explanation of why it do not give exactly zero is the
> > non linearities in the muscle model.
> >
> > Then the question why do the model with the small mass give a
> > larger
> > passive force?
> > When the muscle is loaded with the weight of a small mass, the
> > tendon will be shorter than the calibrated length because in the
> > calibrated length the strain of the tendon for activity one
> > is taken into account. In other words: the muscle is calibrated
for
> >
> > activity one. Since we in both cases drive the overall lengh of
the
> >
> > muscle to the same length the passive element will be more
> > strecthed
> > in the case where we have the small mass, because the tendon is
not
> >
> > loaded so much.
> >
> > So you may ask why did we choose to take into account the tendon
> > strain when we calibrate the muscle?
> > This we did because we wanted to have the optimum strength
exactly
> > at the calibrated length this would not have been the result
> > otherwise.
> >
> > The passive element should work exactly as you describe in the
end
> > of your email, but when the muscle is not loaded very much the
> > passive force will not be zero at the calibrated length.
> > Furthermore
> > the parameters of your model enhance this phenomean because your
> > tendon length is extremely long.
> >
> >
> > I hope this answers your question otherwise please do not
hesitate
> > to write again.
> >
> > Best regards
> > AnyBody Support
> >
> >
> >
> >
> >
> >
> > — In anyscript@yahoogroups.com, “Leng-Feng Lee” <lengfenglee@>
> >
> > wrote:
> > >
> > > Hello,
> > >
> > > I have a muscle modeled using AnyMuscleModel3E, it is connect
> > through
> > > viapointmuscle and the distance between the two points it
> > connects
> > is
> > > 0.5m. I gave the following properties for it:
> > >
> > > AnyMuscleModel3E MuscleModelC = {
> > > F0 = 100;
> > > Lfbar = 0.5;
> > > Lt0 = 0.5;
> > > Epsilonbar = 0.05;
> > > Gammabar = 30*pi/180; //Pennation Angle
> > > Fcfast = 0.4; //Fraction of fast twitch fibers
> > > Jt = 3.0; //Elasiticity of tendon (serial)
> > > Jpe = 3.0; //Elasiticity of tendon
(Parallel)
> > > K1 = 2;
> > > K2 = 8;
> > > PEFactor = 5;
> > > };
> > >
> > > It contract to lift up a segment connected to a joint-no
external
> >
> > load.
> > >
> > > My problem is that the passive force appear throughout the
> > > simulation—something not desired.
> > >
> > > I calibrated the muscle at the initial position, and try to
> > change
> > the
> > > Lfbar and Lt0 value, but the passive force still exist
throughout
> >
> > the
> > > simulation.
> > >
> > > I would like to have a passive force to act in the following
way:
> >
> > at
> > > the initial stage, there is no passive force, as the muscle
> > contract,
> > > the passive force store up(like compressing a spring) and act
in
> > a
> > > direction opposing the contracting force(?), and as the motion
> > change
> > > direction, the passive force release(like let go a compressed
> > spring).
> > >
> > > Could this be done? What parameter(s) should I change/adjust?
Or
> > my
> > > concept about the muscle modeling is wrong? Detail explaination
> > > greately appreciated.
> > >
> > > Thanks,
> > > Leng-Feng
> > >
> >
> >
> >
> >
> >
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
>

Hi,

You said:
“So when the motion is too fast the rigth muscle will be actiated
first because it has to brake the motion and change the direction of
it.”

when you do this:
Freq = 2/Main.OneLegStudy.tEnd ;

and increase tEnd to 2.5sec or 10 sec, this is actually ‘slowing
down’ the motion, right? if tEnd is 1 sec, the motion has to complete
two cycles in one minutes, so it has to move faster. Instead, if the
tEnd is 10 sec, it took 10 sec to complete two cycles–this motion
will be slower. I am confused…

also, my previous question:
And, am I right that the “negative valued” force in the muscle will
shown as ‘zero’ in the plotted force profile? If so, does the anybody
computation machines take the ‘negative valued’ force into account
when it does the subsequence computation or it will take ‘zero’ as a
value?

Another question:
When we calibrate the muscle, is it true that we when we do this, the
spring elements in the muscle is at its rest length? such that when
the muscle is elongated, it create a ‘pulling/contraction’ force
opposing the direction of pull; if the muscle is shorten, it create a
‘repulsive/pushing’ force that oppose the direction of shortening?

Thanks for the time in answering so many questions.

Thanks,
Leng-Feng

— AnyBody Support <support@anybodytech.com> wrote:

> Hi Leng-Feng
>
> Nice little model i have taken a look at it.
>
> There is nothing worng with the muscle parameters, it is the
> dynamics effects of teh kinematics that makes the trick here.
>
> If you change the driver sligthly to be dependent on the tEnd value
>
> of the analysis like this
>
> AnyKinEqFourierDriver Driver = {
> Type = Sin;
> Freq = 2/Main.OneLegStudy.tEnd ; //# of cycles complete
> in 1 sec
> A = {{0,1/2}}; //Control the amplitude of swing
> B = {{0,0}};
> AnyRevoluteJoint &Joint = Main.OneLegModel.Jnts.LegJoint;
> Reaction.Type = {Off};
> };
>
> If gradually increase the tEnd time of the analysis (this is a
> value
> you can set in your study) you will see that about tEnd=2.5 sec the
>
> right muscles statrts to be activated first.When you increase tEnd
> to 10 sec you have the bell shapes you are looking for i belive.
>
> So when the motion is too fast the rigth muscle will be actiated
> first because it has to brake the motion and change the direction
> of
> it.
>
> Our normal muscles can not have pushing forces this not possible in
>
> any way, but the AnyGeneralMuscles can be set to act like a pushing
>
> element, this is done be setting their ForceDirection paramter to
> 1.
>
> Best regards
> AnyBody Support
>
> — In anyscript@yahoogroups.com, Leng-Feng Lee <lengfenglee@…>
> wrote:
> >
> > Hi,
> >
> > Thanks for the quick reply and the demo.
> >
> > To better understand my problem at hand, I have uploaded a model
> call
> > “OneSegModel.any”, which uses two muscles to move a pivoted
> segment
> > swinging back and forth.
> >
> > The muscles are calibrated at its initial position, then run the
> > simulation. After the simulation, the muscle forces Ft/Fm both
> have
> > the similar profile, as the Passive force Fp.
> >
> > It is strange that when the segment swing to the right, this
> movement
> > should be cause by the contraction of the right muscle, but
> instead
> > the muscle force profile of the right muscle is zero at this
> > interval–but when look at muscle force profile of the left
> muscle,
> > it has a bell curve. I wonder why this is so? It act as if
> the ‘swing
> > to the right’ motion is cause (push) by the passive force in the
> Left
> > muscle (?).
> >
> > I guess (?) my muscle model parameters didn’t set up correctly
> such
> > that after the calibration, there are still a lot of passive
> force
> in
> > the muscle? Or what could be the problem?
> >
> > And, am I right that the “negative valued” force in the muscle
> will
> > shown as ‘zero’ in the plotted force profile? If so, does the
> anybody
> > computation machines take the ‘negative valued’ force into
> account
> > when it does the subsequence computation or it will take ‘zero’
> as
> a
> > value?
> >
> > Many thanks,
> > Leng-Feng
> >
> > — AnyBody Support <support@…> wrote:
> >
> > > Hi Leng-Feng
> > >
> > > I have uploaded a small example named
> “MuscleModel3eLengFeng.any”
> > > to
> > > the file section of the group.
> > >
> > > It is a small model to demonstrate the 3 element hill muscle
> model.
> > > The model consist of one mass which is attached to the global
> ref
> > > using a prismatic joint the mass is hanging in a muscle which
> will
> > > lift the mass upward when running the anlysis.
> > >
> > > To run the model please calibrate first then run the inverse
> anal.
> > >
> > > In the current state of the model the mass of the segment is 6
> kg
> > > This will give activity slightly above one
> > > The passive force will have a value in the two first steps but
> it
> > > is
> > > very small
> > > time step1 Fp=0.0155 N
> > > time step2 Fp=0.0036 N
> > > time step3 Fp=0
> > >
> > > If you now try to change the mass of the segment to almost zero
>
> > > Mass=0.05 kg
> > > Then rerun the analysis you will get these results for the
> first
> > > steps
> > > Activity = 0.053
> > > Passive Force Fp=0.131
> > > The passive force will continue to have values until timestep
> 11
> > >
> > > It may seem strange that the passive force is biggest in the
> case
> > > with the small mass, here is the explanation:
> > >
> > > When the calibration is preformed we optimize the muscle to
> have
> > > its
> > > optimal strength at the length in which it is calibrated. We
> simply
> > >
> > > take the overall length from origin to insertation and subtract
>
> the
> > >
> > > lenght of the contractile element to obtain the tendon length.
> > > However we also take into account the strain of the muscle when
>
> it
> > > is maximally loaded, when we do the calibration.
> > > This means that teoretically the passive force should be zero
> when
> > > the activity is one, this is also what the results of the
> analysis
> > > shows. The explanation of why it do not give exactly zero is
> the
> > > non linearities in the muscle model.
> > >
> > > Then the question why do the model with the small mass give a
> > > larger
> > > passive force?
> > > When the muscle is loaded with the weight of a small mass, the
> > > tendon will be shorter than the calibrated length because in
> the
> > > calibrated length the strain of the tendon for activity one
> > > is taken into account. In other words: the muscle is calibrated
>
> for
> > >
> > > activity one. Since we in both cases drive the overall lengh of
>
> the
> > >
> > > muscle to the same length the passive element will be more
> > > strecthed
> > > in the case where we have the small mass, because the tendon is
>
> not
> > >
> > > loaded so much.
> > >
> > > So you may ask why did we choose to take into account the
> tendon
> > > strain when we calibrate the muscle?
> > > This we did because we wanted to have the optimum strength
> exactly
> > > at the calibrated length this would not have been the result
> > > otherwise.
> > >
> > > The passive element should work exactly as you describe in the
> end
> > > of your email, but when the muscle is not loaded very much the
> > > passive force will not be zero at the calibrated length.
> > > Furthermore
> > > the parameters of your model enhance this phenomean because
> your
> > > tendon length is extremely long.
> > >
> > >
> > > I hope this answers your question otherwise please do not
> hesitate
> > > to write again.
> > >
> > > Best regards
> > > AnyBody Support
> > >
> > >
> > >
> > >
> > >
> > >
> > > — In anyscript@yahoogroups.com, “Leng-Feng Lee”
> <lengfenglee@>
> > >
> > > wrote:
> > > >
> > > > Hello,
> > > >
> > > > I have a muscle modeled using AnyMuscleModel3E, it is connect
>
> > > through
> > > > viapointmuscle and the distance between the two points it
> > > connects
> > > is
> > > > 0.5m. I gave the following properties for it:
> > > >
> > > > AnyMuscleModel3E MuscleModelC = {
> > > > F0 = 100;
> > > > Lfbar = 0.5;
> > > > Lt0 = 0.5;
> > > > Epsilonbar = 0.05;
> > > > Gammabar = 30*pi/180; //Pennation Angle
> > > > Fcfast = 0.4; //Fraction of fast twitch
> fibers
> > > > Jt = 3.0; //Elasiticity of tendon
> (serial)
> > > > Jpe = 3.0; //Elasiticity of tendon
> (Parallel)
> > > > K1 = 2;
> > > > K2 = 8;
> > > > PEFactor = 5;
> > > > };
> > > >
> > > > It contract to lift up a segment connected to a joint-no
> external
> > >
> > > load.
> > > >
> > > > My problem is that the passive force appear throughout the
> > > > simulation—something not desired.
> > > >
> > > > I calibrated the muscle at the initial position, and try to
> > > change
> > > the
> > > > Lfbar and Lt0 value, but the passive force still exist
> throughout
> > >
> > > the
> > > > simulation.
> > > >
> > > > I would like to have a passive force to act in the following
> way:
> > >
> > > at
> > > > the initial stage, there is no passive force, as the muscle
> > > contract,
> > > > the passive force store up(like compressing a spring) and act
>
> in
> > > a
> > > > direction opposing the contracting force(?), and as the
> motion
> > > change
> > > > direction, the passive force release(like let go a compressed
>
> > > spring).
> > > >
> > > > Could this be done? What parameter(s) should I change/adjust?
>
> Or
> > > my
> > > > concept about the muscle modeling is wrong? Detail
> explaination
> > > > greately appreciated.
> > > >
> > > > Thanks,
> > > > Leng-Feng
> > > >
> > >
> > >
> > >
> > >
> > >
> >
> >
> > __________________________________________________
> > Do You Yahoo!?
> > Tired of spam? Yahoo! Mail has the best spam protection around
> > http://mail.yahoo.com
> >
>
>
>
>
>
>


Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com

Hi Leng-Feng

I think i messed up right and left in that comment it should have
been left instead of right there. I also wrote

“If gradually increase the tEnd time of the analysis (this is a
value you can set in your study) you will see that about tEnd=2.5
sec the right muscles starts to be activated first”

It is correct that by increasing tEnd the motion cn be slowed down.

So in short slowing down the motion by increasing tEnd removes the
dynamic effects of this problem and the rigth muscles will be
activated first as expected for a slow motion.

The normal muscles in AnyBody can not produce any negative force
(push), so there is no negative force appearing in the muscle
calcualtion at any time, there is no output which are set to zero.
If the muscle can not reduce the overall activity level of the model
it will not be activated, so it is either positive or zero.

It is another story about the AnyGeneralMuscles these can be
configured to be able to push.

When the muscle is calibrated the tendon length is adjusted so that
it will have its maximum strength at that particular postion. This
is done by looking at the overall length of the muscle and
subtracting the contractile element length to obtain the tendon
length. So even when the muscle is elongated at zero activity it
will build up a force in the tendon and passive element. When the
muscle is shorter than the calibrated length, it will not build up
any pushing force at zero activity, it will just be relaxed.

Best regards
AnyBody Support

— In anyscript@yahoogroups.com, Leng-Feng Lee <lengfenglee@…>
wrote:
>
> Hi,
>
> You said:
> “So when the motion is too fast the rigth muscle will be actiated
> first because it has to brake the motion and change the direction
of
> it.”
>
> when you do this:
> Freq = 2/Main.OneLegStudy.tEnd ;
>
> and increase tEnd to 2.5sec or 10 sec, this is actually ‘slowing
> down’ the motion, right? if tEnd is 1 sec, the motion has to
complete
> two cycles in one minutes, so it has to move faster. Instead, if
the
> tEnd is 10 sec, it took 10 sec to complete two cycles–this motion
> will be slower. I am confused…
>
> also, my previous question:
> And, am I right that the “negative valued” force in the muscle will
> shown as ‘zero’ in the plotted force profile? If so, does the
anybody
> computation machines take the ‘negative valued’ force into account
> when it does the subsequence computation or it will take ‘zero’
as a
> value?
>
> Another question:
> When we calibrate the muscle, is it true that we when we do this,
the
> spring elements in the muscle is at its rest length? such that when
> the muscle is elongated, it create a ‘pulling/contraction’ force
> opposing the direction of pull; if the muscle is shorten, it
create a
> ‘repulsive/pushing’ force that oppose the direction of shortening?
>
> Thanks for the time in answering so many questions.
>
> Thanks,
> Leng-Feng
>
>
> — AnyBody Support <support@…> wrote:
>
> > Hi Leng-Feng
> >
> > Nice little model i have taken a look at it.
> >
> > There is nothing worng with the muscle parameters, it is the
> > dynamics effects of teh kinematics that makes the trick here.
> >
> > If you change the driver sligthly to be dependent on the tEnd
value
> >
> > of the analysis like this
> >
> > AnyKinEqFourierDriver Driver = {
> > Type = Sin;
> > Freq = 2/Main.OneLegStudy.tEnd ; //# of cycles
complete
> > in 1 sec
> > A = {{0,1/2}}; //Control the amplitude of swing
> > B = {{0,0}};
> > AnyRevoluteJoint &Joint = Main.OneLegModel.Jnts.LegJoint;
> > Reaction.Type = {Off};
> > };
> >
> > If gradually increase the tEnd time of the analysis (this is a
> > value
> > you can set in your study) you will see that about tEnd=2.5 sec
the
> >
> > right muscles statrts to be activated first.When you increase
tEnd
> > to 10 sec you have the bell shapes you are looking for i belive.
> >
> > So when the motion is too fast the rigth muscle will be actiated
> > first because it has to brake the motion and change the direction
> > of
> > it.
> >
> > Our normal muscles can not have pushing forces this not possible
in
> >
> > any way, but the AnyGeneralMuscles can be set to act like a
pushing
> >
> > element, this is done be setting their ForceDirection paramter to
> > 1.
> >
> > Best regards
> > AnyBody Support
> >
> > — In anyscript@yahoogroups.com, Leng-Feng Lee <lengfenglee@>
> > wrote:
> > >
> > > Hi,
> > >
> > > Thanks for the quick reply and the demo.
> > >
> > > To better understand my problem at hand, I have uploaded a
model
> > call
> > > “OneSegModel.any”, which uses two muscles to move a pivoted
> > segment
> > > swinging back and forth.
> > >
> > > The muscles are calibrated at its initial position, then run
the
> > > simulation. After the simulation, the muscle forces Ft/Fm both
> > have
> > > the similar profile, as the Passive force Fp.
> > >
> > > It is strange that when the segment swing to the right, this
> > movement
> > > should be cause by the contraction of the right muscle, but
> > instead
> > > the muscle force profile of the right muscle is zero at this
> > > interval–but when look at muscle force profile of the left
> > muscle,
> > > it has a bell curve. I wonder why this is so? It act as if
> > the ‘swing
> > > to the right’ motion is cause (push) by the passive force in
the
> > Left
> > > muscle (?).
> > >
> > > I guess (?) my muscle model parameters didn’t set up correctly
> > such
> > > that after the calibration, there are still a lot of passive
> > force
> > in
> > > the muscle? Or what could be the problem?
> > >
> > > And, am I right that the “negative valued” force in the muscle
> > will
> > > shown as ‘zero’ in the plotted force profile? If so, does the
> > anybody
> > > computation machines take the ‘negative valued’ force into
> > account
> > > when it does the subsequence computation or it will take ‘zero’
> > as
> > a
> > > value?
> > >
> > > Many thanks,
> > > Leng-Feng
> > >
> > > — AnyBody Support <support@> wrote:
> > >
> > > > Hi Leng-Feng
> > > >
> > > > I have uploaded a small example named
> > “MuscleModel3eLengFeng.any”
> > > > to
> > > > the file section of the group.
> > > >
> > > > It is a small model to demonstrate the 3 element hill muscle
> > model.
> > > > The model consist of one mass which is attached to the
global
> > ref
> > > > using a prismatic joint the mass is hanging in a muscle
which
> > will
> > > > lift the mass upward when running the anlysis.
> > > >
> > > > To run the model please calibrate first then run the inverse
> > anal.
> > > >
> > > > In the current state of the model the mass of the segment is
6
> > kg
> > > > This will give activity slightly above one
> > > > The passive force will have a value in the two first steps
but
> > it
> > > > is
> > > > very small
> > > > time step1 Fp=0.0155 N
> > > > time step2 Fp=0.0036 N
> > > > time step3 Fp=0
> > > >
> > > > If you now try to change the mass of the segment to almost
zero
> >
> > > > Mass=0.05 kg
> > > > Then rerun the analysis you will get these results for the
> > first
> > > > steps
> > > > Activity = 0.053
> > > > Passive Force Fp=0.131
> > > > The passive force will continue to have values until timestep
> > 11
> > > >
> > > > It may seem strange that the passive force is biggest in the
> > case
> > > > with the small mass, here is the explanation:
> > > >
> > > > When the calibration is preformed we optimize the muscle to
> > have
> > > > its
> > > > optimal strength at the length in which it is calibrated. We
> > simply
> > > >
> > > > take the overall length from origin to insertation and
subtract
> >
> > the
> > > >
> > > > lenght of the contractile element to obtain the tendon
length.
> > > > However we also take into account the strain of the muscle
when
> >
> > it
> > > > is maximally loaded, when we do the calibration.
> > > > This means that teoretically the passive force should be
zero
> > when
> > > > the activity is one, this is also what the results of the
> > analysis
> > > > shows. The explanation of why it do not give exactly zero is
> > the
> > > > non linearities in the muscle model.
> > > >
> > > > Then the question why do the model with the small mass give a
> > > > larger
> > > > passive force?
> > > > When the muscle is loaded with the weight of a small mass,
the
> > > > tendon will be shorter than the calibrated length because in
> > the
> > > > calibrated length the strain of the tendon for activity one
> > > > is taken into account. In other words: the muscle is
calibrated
> >
> > for
> > > >
> > > > activity one. Since we in both cases drive the overall lengh
of
> >
> > the
> > > >
> > > > muscle to the same length the passive element will be more
> > > > strecthed
> > > > in the case where we have the small mass, because the tendon
is
> >
> > not
> > > >
> > > > loaded so much.
> > > >
> > > > So you may ask why did we choose to take into account the
> > tendon
> > > > strain when we calibrate the muscle?
> > > > This we did because we wanted to have the optimum strength
> > exactly
> > > > at the calibrated length this would not have been the result
> > > > otherwise.
> > > >
> > > > The passive element should work exactly as you describe in
the
> > end
> > > > of your email, but when the muscle is not loaded very much
the
> > > > passive force will not be zero at the calibrated length.
> > > > Furthermore
> > > > the parameters of your model enhance this phenomean because
> > your
> > > > tendon length is extremely long.
> > > >
> > > >
> > > > I hope this answers your question otherwise please do not
> > hesitate
> > > > to write again.
> > > >
> > > > Best regards
> > > > AnyBody Support
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > > — In anyscript@yahoogroups.com, “Leng-Feng Lee”
> > <lengfenglee@>
> > > >
> > > > wrote:
> > > > >
> > > > > Hello,
> > > > >
> > > > > I have a muscle modeled using AnyMuscleModel3E, it is
connect
> >
> > > > through
> > > > > viapointmuscle and the distance between the two points it
> > > > connects
> > > > is
> > > > > 0.5m. I gave the following properties for it:
> > > > >
> > > > > AnyMuscleModel3E MuscleModelC = {
> > > > > F0 = 100;
> > > > > Lfbar = 0.5;
> > > > > Lt0 = 0.5;
> > > > > Epsilonbar = 0.05;
> > > > > Gammabar = 30*pi/180; //Pennation Angle
> > > > > Fcfast = 0.4; //Fraction of fast twitch
> > fibers
> > > > > Jt = 3.0; //Elasiticity of tendon
> > (serial)
> > > > > Jpe = 3.0; //Elasiticity of tendon
> > (Parallel)
> > > > > K1 = 2;
> > > > > K2 = 8;
> > > > > PEFactor = 5;
> > > > > };
> > > > >
> > > > > It contract to lift up a segment connected to a joint-no
> > external
> > > >
> > > > load.
> > > > >
> > > > > My problem is that the passive force appear throughout the
> > > > > simulation—something not desired.
> > > > >
> > > > > I calibrated the muscle at the initial position, and try to
> > > > change
> > > > the
> > > > > Lfbar and Lt0 value, but the passive force still exist
> > throughout
> > > >
> > > > the
> > > > > simulation.
> > > > >
> > > > > I would like to have a passive force to act in the
following
> > way:
> > > >
> > > > at
> > > > > the initial stage, there is no passive force, as the
muscle
> > > > contract,
> > > > > the passive force store up(like compressing a spring) and
act
> >
> > in
> > > > a
> > > > > direction opposing the contracting force(?), and as the
> > motion
> > > > change
> > > > > direction, the passive force release(like let go a
compressed
> >
> > > > spring).
> > > > >
> > > > > Could this be done? What parameter(s) should I
change/adjust?
> >
> > Or
> > > > my
> > > > > concept about the muscle modeling is wrong? Detail
> > explaination
> > > > > greately appreciated.
> > > > >
> > > > > Thanks,
> > > > > Leng-Feng
> > > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > > __________________________________________________
> > > Do You Yahoo!?
> > > Tired of spam? Yahoo! Mail has the best spam protection
around
> > > http://mail.yahoo.com
> > >
> >
> >
> >
> >
> >
> >
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
>

Hi,

Thanks for the explaination.

I still don’t quite get this:
“So in short slowing down the motion by increasing tEnd removes the
dynamic effects of this problem and the rigth muscles will be
activated first as expected for a slow motion.”

–when the segment swing to the right, shouldn’t this motion is cause
by the contraction of the right muscle (no matter how fast it move)?
why when it is moving fast, the right muscle is not being activate
first (in this case, what causes the movement?)?

Another question regarding this:
" The normal muscles in AnyBody can not produce any negative force
(push), so there is no negative force appearing in the muscle
calcualtion at any time, there is no output which are set to zero. If
the muscle can not reduce the overall activity level of the model it
will not be activated, so it is either positive or zero."

–How about the passive force? will the built-up passive force in the
right muscle during its swinging motion to the right helps to ‘push’
the segment to the left when the motion is going to the left (like
the compressed spring does when it was release)?

My assumption is this:
During the ‘swinging right motion’, the contractile muscle in the
right mucle contracts, producing a force to move the segment–this
force also need to overcome the built-up passive force in the right
mucle(due to contraction), and also the built-up passive force in the
left muscle (due to the elongation of the left muscle). When the
segment reach the right most position and starts to swing to the left
, the left muscle contract–but it doesn’t need that much of force
because the gravity, the passive force in the right and left muscles
all contribute to this left swinging motion at that instant. The left
muscle will need more force when the segment pass throught the bottom
position: since now the passive force in both muscle and the gravity
force act in the opposite direction of the left muscle contraction
force.

–Am I right about this? Please let me know if my understanding is
wrong–that muscle doesn’t act in the manner I explain. Or if AnyBody
didn’t take into account of it?

Thank you very much in your time!

Thanks,
Leng-Feng

— AnyBody Support <support@anybodytech.com> wrote:

> Hi Leng-Feng
>
> I think i messed up right and left in that comment it should have
> been left instead of right there. I also wrote
>
> “If gradually increase the tEnd time of the analysis (this is a
> value you can set in your study) you will see that about tEnd=2.5
> sec the right muscles starts to be activated first”
>
> It is correct that by increasing tEnd the motion cn be slowed down.
>
> So in short slowing down the motion by increasing tEnd removes the
> dynamic effects of this problem and the rigth muscles will be
> activated first as expected for a slow motion.
>
> The normal muscles in AnyBody can not produce any negative force
> (push), so there is no negative force appearing in the muscle
> calcualtion at any time, there is no output which are set to zero.
> If the muscle can not reduce the overall activity level of the
> model
> it will not be activated, so it is either positive or zero.
>
> It is another story about the AnyGeneralMuscles these can be
> configured to be able to push.
>
> When the muscle is calibrated the tendon length is adjusted so that
>
> it will have its maximum strength at that particular postion. This
> is done by looking at the overall length of the muscle and
> subtracting the contractile element length to obtain the tendon
> length. So even when the muscle is elongated at zero activity it
> will build up a force in the tendon and passive element. When the
> muscle is shorter than the calibrated length, it will not build up
> any pushing force at zero activity, it will just be relaxed.
>
>
> Best regards
> AnyBody Support
>
> — In anyscript@yahoogroups.com, Leng-Feng Lee <lengfenglee@…>
> wrote:
> >
> > Hi,
> >
> > You said:
> > “So when the motion is too fast the rigth muscle will be actiated
> > first because it has to brake the motion and change the direction
>
> of
> > it.”
> >
> > when you do this:
> > Freq = 2/Main.OneLegStudy.tEnd ;
> >
> > and increase tEnd to 2.5sec or 10 sec, this is actually ‘slowing
> > down’ the motion, right? if tEnd is 1 sec, the motion has to
> complete
> > two cycles in one minutes, so it has to move faster. Instead, if
> the
> > tEnd is 10 sec, it took 10 sec to complete two cycles–this
> motion
> > will be slower. I am confused…
> >
> > also, my previous question:
> > And, am I right that the “negative valued” force in the muscle
> will
> > shown as ‘zero’ in the plotted force profile? If so, does the
> anybody
> > computation machines take the ‘negative valued’ force into
> account
> > when it does the subsequence computation or it will take ‘zero’
> as a
> > value?
> >
> > Another question:
> > When we calibrate the muscle, is it true that we when we do this,
>
> the
> > spring elements in the muscle is at its rest length? such that
> when
> > the muscle is elongated, it create a ‘pulling/contraction’ force
> > opposing the direction of pull; if the muscle is shorten, it
> create a
> > ‘repulsive/pushing’ force that oppose the direction of
> shortening?
> >
> > Thanks for the time in answering so many questions.
> >
> > Thanks,
> > Leng-Feng
> >
> >
> > — AnyBody Support <support@…> wrote:
> >
> > > Hi Leng-Feng
> > >
> > > Nice little model i have taken a look at it.
> > >
> > > There is nothing worng with the muscle parameters, it is the
> > > dynamics effects of teh kinematics that makes the trick here.
> > >
> > > If you change the driver sligthly to be dependent on the tEnd
> value
> > >
> > > of the analysis like this
> > >
> > > AnyKinEqFourierDriver Driver = {
> > > Type = Sin;
> > > Freq = 2/Main.OneLegStudy.tEnd ; //# of cycles
> complete
> > > in 1 sec
> > > A = {{0,1/2}}; //Control the amplitude of swing
> > > B = {{0,0}};
> > > AnyRevoluteJoint &Joint = Main.OneLegModel.Jnts.LegJoint;
> > > Reaction.Type = {Off};
> > > };
> > >
> > > If gradually increase the tEnd time of the analysis (this is a
> > > value
> > > you can set in your study) you will see that about tEnd=2.5 sec
>
> the
> > >
> > > right muscles statrts to be activated first.When you increase
> tEnd
> > > to 10 sec you have the bell shapes you are looking for i
> belive.
> > >
> > > So when the motion is too fast the rigth muscle will be
> actiated
> > > first because it has to brake the motion and change the
> direction
> > > of
> > > it.
> > >
> > > Our normal muscles can not have pushing forces this not
> possible
> in
> > >
> > > any way, but the AnyGeneralMuscles can be set to act like a
> pushing
> > >
> > > element, this is done be setting their ForceDirection paramter
> to
> > > 1.
> > >
> > > Best regards
> > > AnyBody Support
> > >
> > > — In anyscript@yahoogroups.com, Leng-Feng Lee <lengfenglee@>
> > > wrote:
> > > >
> > > > Hi,
> > > >
> > > > Thanks for the quick reply and the demo.
> > > >
> > > > To better understand my problem at hand, I have uploaded a
> model
> > > call
> > > > “OneSegModel.any”, which uses two muscles to move a pivoted
> > > segment
> > > > swinging back and forth.
> > > >
> > > > The muscles are calibrated at its initial position, then run
> the
> > > > simulation. After the simulation, the muscle forces Ft/Fm
> both
> > > have
> > > > the similar profile, as the Passive force Fp.
> > > >
> > > > It is strange that when the segment swing to the right, this
> > > movement
> > > > should be cause by the contraction of the right muscle, but
> > > instead
> > > > the muscle force profile of the right muscle is zero at this
> > > > interval–but when look at muscle force profile of the left
> > > muscle,
> > > > it has a bell curve. I wonder why this is so? It act as if
> > > the ‘swing
> > > > to the right’ motion is cause (push) by the passive force in
> the
> > > Left
> > > > muscle (?).
> > > >
> > > > I guess (?) my muscle model parameters didn’t set up
> correctly
> > > such
> > > > that after the calibration, there are still a lot of passive
> > > force
> > > in
> > > > the muscle? Or what could be the problem?
> > > >
> > > > And, am I right that the “negative valued” force in the
> muscle
> > > will
> > > > shown as ‘zero’ in the plotted force profile? If so, does the
>
> > > anybody
> > > > computation machines take the ‘negative valued’ force into
> > > account
> > > > when it does the subsequence computation or it will take
> ‘zero’
> > > as
> > > a
> > > > value?
> > > >
> > > > Many thanks,
> > > > Leng-Feng
> > > >
> > > > — AnyBody Support <support@> wrote:
> > > >
> > > > > Hi Leng-Feng
> > > > >
> > > > > I have uploaded a small example named
> > > “MuscleModel3eLengFeng.any”
> > > > > to
> > > > > the file section of the group.
> > > > >
> > > > > It is a small model to demonstrate the 3 element hill
> muscle
> > > model.
> > > > > The model consist of one mass which is attached to the
> global
> > > ref
> > > > > using a prismatic joint the mass is hanging in a muscle
> which
> > > will
> > > > > lift the mass upward when running the anlysis.
> > > > >
> > > > > To run the model please calibrate first then run the
> inverse
> > > anal.
> > > > >
> > > > > In the current state of the model the mass of the segment
> is
> 6
> > > kg
> > > > > This will give activity slightly above one
> > > > > The passive force will have a value in the two first steps
> but
> > > it
> > > > > is
> > > > > very small
> > > > > time step1 Fp=0.0155 N
> > > > > time step2 Fp=0.0036 N
> > > > > time step3 Fp=0
> > > > >
> > > > > If you now try to change the mass of the segment to almost
> zero
> > >
> > > > > Mass=0.05 kg
> > > > > Then rerun the analysis you will get these results for the
> > > first
> > > > > steps
> > > > > Activity = 0.053
> > > > > Passive Force Fp=0.131
> > > > > The passive force will continue to have values until
> timestep
> > > 11
> > > > >
> > > > > It may seem strange that the passive force is biggest in
> the
> > > case
> > > > > with the small mass, here is the explanation:
> > > > >
> > > > > When the calibration is preformed we optimize the muscle to
> > > have
> > > > > its
> > > > > optimal strength at the length in which it is calibrated.
> We
> > > simply
> > > > >
> > > > > take the overall length from origin to insertation and
> subtract
> > >
> > > the
> > > > >
> > > > > lenght of the contractile element to obtain the tendon
> length.
> > > > > However we also take into account the strain of the muscle
> when
> > >
> > > it
> > > > > is maximally loaded, when we do the calibration.
> > > > > This means that teoretically the passive force should be
> zero
> > > when
> > > > > the activity is one, this is also what the results of the
> > > analysis
> > > > > shows. The explanation of why it do not give exactly zero
> is
> > > the
> > > > > non linearities in the muscle model.
> > > > >
> > > > > Then the question why do the model with the small mass give
> a
> > > > > larger
> > > > > passive force?
> > > > > When the muscle is loaded with the weight of a small mass,
> the
> > > > > tendon will be shorter than the calibrated length because
> in
> > > the
> > > > > calibrated length the strain of the tendon for activity one
> > > > > is taken into account. In other words: the muscle is
> calibrated
> > >
> > > for
> > > > >
> > > > > activity one. Since we in both cases drive the overall
> lengh
> of
> > >
> > > the
> > > > >
> > > > > muscle to the same length the passive element will be more
> > > > > strecthed
> > > > > in the case where we have the small mass, because the
> tendon
> is
> > >
> > > not
> > > > >
> > > > > loaded so much.
> > > > >
> > > > > So you may ask why did we choose to take into account the
> > > tendon
> > > > > strain when we calibrate the muscle?
> > > > > This we did because we wanted to have the optimum strength
> > > exactly
> > > > > at the calibrated length this would not have been the
> result
> > > > > otherwise.
> > > > >
> > > > > The passive element should work exactly as you describe in
> the
> > > end
> > > > > of your email, but when the muscle is not loaded very much
> the
> > > > > passive force will not be zero at the calibrated length.
> > > > > Furthermore
> > > > > the parameters of your model enhance this phenomean because
> > > your
> > > > > tendon length is extremely long.
> > > > >
> > > > >
> > > > > I hope this answers your question otherwise please do not
> > > hesitate
> > > > > to write again.
> > > > >
> > > > > Best regards
> > > > > AnyBody Support
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > — In anyscript@yahoogroups.com, “Leng-Feng Lee”
> > > <lengfenglee@>
> > > > >
> > > > > wrote:
> > > > > >
> > > > > > Hello,
> > > > > >
> > > > > > I have a muscle modeled using AnyMuscleModel3E, it is
> connect
> > >
> > > > > through
> > > > > > viapointmuscle and the distance between the two points it
> > > > > connects
> > > > > is
> > > > > > 0.5m. I gave the following properties for it:
> > > > > >
> > > > > > AnyMuscleModel3E MuscleModelC = {
> > > > > > F0 = 100;
> > > > > > Lfbar = 0.5;
> > > > > > Lt0 = 0.5;
> > > > > > Epsilonbar = 0.05;
> > > > > > Gammabar = 30*pi/180; //Pennation Angle
> > > > > > Fcfast = 0.4; //Fraction of fast twitch
> > > fibers
> > > > > > Jt = 3.0; //Elasiticity of tendon
> > > (serial)
> > > > > > Jpe = 3.0; //Elasiticity of tendon
> > > (Parallel)
> > > > > > K1 = 2;
> > > > > > K2 = 8;
> > > > > > PEFactor = 5;
> > > > > > };
> > > > > >
> > > > > > It contract to lift up a segment connected to a joint-no
> > > external
> > > > >
> > > > > load.
> > > > > >
> > > > > > My problem is that the passive force appear throughout
> the
> > > > > > simulation—something not desired.
> > > > > >
> > > > > > I calibrated the muscle at the initial position, and try
> to
> > > > > change
> > > > > the
> > > > > > Lfbar and Lt0 value, but the passive force still exist
> > > throughout
> > > > >
> > > > > the
> > > > > > simulation.
> > > > > >
> > > > > > I would like to have a passive force to act in the
> following
> > > way:
> > > > >
> > > > > at
> > > > > > the initial stage, there is no passive force, as the
> muscle
> > > > > contract,
> > > > > > the passive force store up(like compressing a spring) and
>
> act
> > >
> > > in
> > > > > a
> > > > > > direction opposing the contracting force(?), and as the
> > > motion
> > > > > change
> > > > > > direction, the passive force release(like let go a
> compressed
> > >
> > > > > spring).
> > > > > >
> > > > > > Could this be done? What parameter(s) should I
> change/adjust?
> > >
> > > Or
> > > > > my
> > > > > > concept about the muscle modeling is wrong? Detail
> > > explaination
> > > > > > greately appreciated.
> > > > > >
> > > > > > Thanks,
> > > > > > Leng-Feng
> > > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > >
> > > >
> > > > __________________________________________________
> > > > Do You Yahoo!?
> > > > Tired of spam? Yahoo! Mail has the best spam protection
> around
> > > > http://mail.yahoo.com
> > > >
> > >
> > >
> > >
> > >
> > >
> > >
> >
> >
> > __________________________________________________
> > Do You Yahoo!?
> > Tired of spam? Yahoo! Mail has the best spam protection around
> > http://mail.yahoo.com
> >
>
>
>
>
>


Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com

Hi Leng-Feng

When the motion is slow the muscle to the right will have to be
active to “lift” the pendul. When the motion is fast the left muscle
will have to be activated in order to brake the motion before it
swings the other way. If you inrease the number of timesteps the
motion will appear slower in the modelview window and it is easier
to see this. So yes the activation will depend on how fast the
motion is.

The passive forces can not act like compressive springs there can
not be any pushing forces in the muscles at any time. When the
pendul moves to the right the muscle on the left will build up a
passive force due to its elongation, this force can be used when the
pendul has to swing back.

Your assumptions on how it works are not entirely correct:
For a slow motion when the pendul moves to the right it will build
up passive forces in the left muscle not in the right muscle. So
when it swing back the gravity, the pasive force on the left muscle
and the posible some activity of the rigth muscle to hold back the
motion will work on the pendul, not any compression force from the
right muscle. When the pendul reached the bottom and moves to the
left the left muscle should be active and it start working as the
right muscles did when the pendul moved to the rigth

Hope it is more clear now, otherwise please write again

Best regards
AnyBody Support

— In anyscript@yahoogroups.com, Leng-Feng Lee <lengfenglee@…>
wrote:
>
> Hi,
>
> Thanks for the explaination.
>
> I still don’t quite get this:
> “So in short slowing down the motion by increasing tEnd removes
the
> dynamic effects of this problem and the rigth muscles will be
> activated first as expected for a slow motion.”
>
> --when the segment swing to the right, shouldn’t this motion is
cause
> by the contraction of the right muscle (no matter how fast it
move)?
> why when it is moving fast, the right muscle is not being activate
> first (in this case, what causes the movement?)?
>
> Another question regarding this:
> " The normal muscles in AnyBody can not produce any negative force
> (push), so there is no negative force appearing in the muscle
> calcualtion at any time, there is no output which are set to zero.
If
> the muscle can not reduce the overall activity level of the model
it
> will not be activated, so it is either positive or zero."
>
> --How about the passive force? will the built-up passive force in
the
> right muscle during its swinging motion to the right helps
to ‘push’
> the segment to the left when the motion is going to the left (like
> the compressed spring does when it was release)?
>
> My assumption is this:
> During the ‘swinging right motion’, the contractile muscle in the
> right mucle contracts, producing a force to move the segment–this
> force also need to overcome the built-up passive force in the right
> mucle(due to contraction), and also the built-up passive force in
the
> left muscle (due to the elongation of the left muscle). When the
> segment reach the right most position and starts to swing to the
left
> , the left muscle contract–but it doesn’t need that much of force
> because the gravity, the passive force in the right and left
muscles
> all contribute to this left swinging motion at that instant. The
left
> muscle will need more force when the segment pass throught the
bottom
> position: since now the passive force in both muscle and the
gravity
> force act in the opposite direction of the left muscle contraction
> force.
>
> --Am I right about this? Please let me know if my understanding is
> wrong–that muscle doesn’t act in the manner I explain. Or if
AnyBody
> didn’t take into account of it?
>
> Thank you very much in your time!
>
>
> Thanks,
> Leng-Feng
>
> — AnyBody Support <support@…> wrote:
>
> > Hi Leng-Feng
> >
> > I think i messed up right and left in that comment it should
have
> > been left instead of right there. I also wrote
> >
> > “If gradually increase the tEnd time of the analysis (this is a
> > value you can set in your study) you will see that about
tEnd=2.5
> > sec the right muscles starts to be activated first”
> >
> > It is correct that by increasing tEnd the motion cn be slowed
down.
> >
> > So in short slowing down the motion by increasing tEnd removes
the
> > dynamic effects of this problem and the rigth muscles will be
> > activated first as expected for a slow motion.
> >
> > The normal muscles in AnyBody can not produce any negative force
> > (push), so there is no negative force appearing in the muscle
> > calcualtion at any time, there is no output which are set to
zero.
> > If the muscle can not reduce the overall activity level of the
> > model
> > it will not be activated, so it is either positive or zero.
> >
> > It is another story about the AnyGeneralMuscles these can be
> > configured to be able to push.
> >
> > When the muscle is calibrated the tendon length is adjusted so
that
> >
> > it will have its maximum strength at that particular postion.
This
> > is done by looking at the overall length of the muscle and
> > subtracting the contractile element length to obtain the tendon
> > length. So even when the muscle is elongated at zero activity it
> > will build up a force in the tendon and passive element. When
the
> > muscle is shorter than the calibrated length, it will not build
up
> > any pushing force at zero activity, it will just be relaxed.
> >
> >
> > Best regards
> > AnyBody Support
> >
> > — In anyscript@yahoogroups.com, Leng-Feng Lee <lengfenglee@>
> > wrote:
> > >
> > > Hi,
> > >
> > > You said:
> > > “So when the motion is too fast the rigth muscle will be
actiated
> > > first because it has to brake the motion and change the
direction
> >
> > of
> > > it.”
> > >
> > > when you do this:
> > > Freq = 2/Main.OneLegStudy.tEnd ;
> > >
> > > and increase tEnd to 2.5sec or 10 sec, this is
actually ‘slowing
> > > down’ the motion, right? if tEnd is 1 sec, the motion has to
> > complete
> > > two cycles in one minutes, so it has to move faster. Instead,
if
> > the
> > > tEnd is 10 sec, it took 10 sec to complete two cycles–this
> > motion
> > > will be slower. I am confused…
> > >
> > > also, my previous question:
> > > And, am I right that the “negative valued” force in the muscle
> > will
> > > shown as ‘zero’ in the plotted force profile? If so, does the
> > anybody
> > > computation machines take the ‘negative valued’ force into
> > account
> > > when it does the subsequence computation or it will
take ‘zero’
> > as a
> > > value?
> > >
> > > Another question:
> > > When we calibrate the muscle, is it true that we when we do
this,
> >
> > the
> > > spring elements in the muscle is at its rest length? such that
> > when
> > > the muscle is elongated, it create a ‘pulling/contraction’
force
> > > opposing the direction of pull; if the muscle is shorten, it
> > create a
> > > ‘repulsive/pushing’ force that oppose the direction of
> > shortening?
> > >
> > > Thanks for the time in answering so many questions.
> > >
> > > Thanks,
> > > Leng-Feng
> > >
> > >
> > > — AnyBody Support <support@> wrote:
> > >
> > > > Hi Leng-Feng
> > > >
> > > > Nice little model i have taken a look at it.
> > > >
> > > > There is nothing worng with the muscle parameters, it is the
> > > > dynamics effects of teh kinematics that makes the trick here.
> > > >
> > > > If you change the driver sligthly to be dependent on the
tEnd
> > value
> > > >
> > > > of the analysis like this
> > > >
> > > > AnyKinEqFourierDriver Driver = {
> > > > Type = Sin;
> > > > Freq = 2/Main.OneLegStudy.tEnd ; //# of cycles
> > complete
> > > > in 1 sec
> > > > A = {{0,1/2}}; //Control the amplitude of swing
> > > > B = {{0,0}};
> > > > AnyRevoluteJoint &Joint =
Main.OneLegModel.Jnts.LegJoint;
> > > > Reaction.Type = {Off};
> > > > };
> > > >
> > > > If gradually increase the tEnd time of the analysis (this is
a
> > > > value
> > > > you can set in your study) you will see that about tEnd=2.5
sec
> >
> > the
> > > >
> > > > right muscles statrts to be activated first.When you
increase
> > tEnd
> > > > to 10 sec you have the bell shapes you are looking for i
> > belive.
> > > >
> > > > So when the motion is too fast the rigth muscle will be
> > actiated
> > > > first because it has to brake the motion and change the
> > direction
> > > > of
> > > > it.
> > > >
> > > > Our normal muscles can not have pushing forces this not
> > possible
> > in
> > > >
> > > > any way, but the AnyGeneralMuscles can be set to act like a
> > pushing
> > > >
> > > > element, this is done be setting their ForceDirection
paramter
> > to
> > > > 1.
> > > >
> > > > Best regards
> > > > AnyBody Support
> > > >
> > > > — In anyscript@yahoogroups.com, Leng-Feng Lee
<lengfenglee@>
> > > > wrote:
> > > > >
> > > > > Hi,
> > > > >
> > > > > Thanks for the quick reply and the demo.
> > > > >
> > > > > To better understand my problem at hand, I have uploaded a
> > model
> > > > call
> > > > > “OneSegModel.any”, which uses two muscles to move a pivoted
> > > > segment
> > > > > swinging back and forth.
> > > > >
> > > > > The muscles are calibrated at its initial position, then
run
> > the
> > > > > simulation. After the simulation, the muscle forces Ft/Fm
> > both
> > > > have
> > > > > the similar profile, as the Passive force Fp.
> > > > >
> > > > > It is strange that when the segment swing to the right,
this
> > > > movement
> > > > > should be cause by the contraction of the right muscle, but
> > > > instead
> > > > > the muscle force profile of the right muscle is zero at
this
> > > > > interval–but when look at muscle force profile of the left
> > > > muscle,
> > > > > it has a bell curve. I wonder why this is so? It act as if
> > > > the ‘swing
> > > > > to the right’ motion is cause (push) by the passive force
in
> > the
> > > > Left
> > > > > muscle (?).
> > > > >
> > > > > I guess (?) my muscle model parameters didn’t set up
> > correctly
> > > > such
> > > > > that after the calibration, there are still a lot of
passive
> > > > force
> > > > in
> > > > > the muscle? Or what could be the problem?
> > > > >
> > > > > And, am I right that the “negative valued” force in the
> > muscle
> > > > will
> > > > > shown as ‘zero’ in the plotted force profile? If so, does
the
> >
> > > > anybody
> > > > > computation machines take the ‘negative valued’ force into
> > > > account
> > > > > when it does the subsequence computation or it will take
> > ‘zero’
> > > > as
> > > > a
> > > > > value?
> > > > >
> > > > > Many thanks,
> > > > > Leng-Feng
> > > > >
> > > > > — AnyBody Support <support@> wrote:
> > > > >
> > > > > > Hi Leng-Feng
> > > > > >
> > > > > > I have uploaded a small example named
> > > > “MuscleModel3eLengFeng.any”
> > > > > > to
> > > > > > the file section of the group.
> > > > > >
> > > > > > It is a small model to demonstrate the 3 element hill
> > muscle
> > > > model.
> > > > > > The model consist of one mass which is attached to the
> > global
> > > > ref
> > > > > > using a prismatic joint the mass is hanging in a muscle
> > which
> > > > will
> > > > > > lift the mass upward when running the anlysis.
> > > > > >
> > > > > > To run the model please calibrate first then run the
> > inverse
> > > > anal.
> > > > > >
> > > > > > In the current state of the model the mass of the segment
> > is
> > 6
> > > > kg
> > > > > > This will give activity slightly above one
> > > > > > The passive force will have a value in the two first
steps
> > but
> > > > it
> > > > > > is
> > > > > > very small
> > > > > > time step1 Fp=0.0155 N
> > > > > > time step2 Fp=0.0036 N
> > > > > > time step3 Fp=0
> > > > > >
> > > > > > If you now try to change the mass of the segment to
almost
> > zero
> > > >
> > > > > > Mass=0.05 kg
> > > > > > Then rerun the analysis you will get these results for
the
> > > > first
> > > > > > steps
> > > > > > Activity = 0.053
> > > > > > Passive Force Fp=0.131
> > > > > > The passive force will continue to have values until
> > timestep
> > > > 11
> > > > > >
> > > > > > It may seem strange that the passive force is biggest in
> > the
> > > > case
> > > > > > with the small mass, here is the explanation:
> > > > > >
> > > > > > When the calibration is preformed we optimize the muscle
to
> > > > have
> > > > > > its
> > > > > > optimal strength at the length in which it is calibrated.
> > We
> > > > simply
> > > > > >
> > > > > > take the overall length from origin to insertation and
> > subtract
> > > >
> > > > the
> > > > > >
> > > > > > lenght of the contractile element to obtain the tendon
> > length.
> > > > > > However we also take into account the strain of the
muscle
> > when
> > > >
> > > > it
> > > > > > is maximally loaded, when we do the calibration.
> > > > > > This means that teoretically the passive force should be
> > zero
> > > > when
> > > > > > the activity is one, this is also what the results of
the
> > > > analysis
> > > > > > shows. The explanation of why it do not give exactly zero
> > is
> > > > the
> > > > > > non linearities in the muscle model.
> > > > > >
> > > > > > Then the question why do the model with the small mass
give
> > a
> > > > > > larger
> > > > > > passive force?
> > > > > > When the muscle is loaded with the weight of a small
mass,
> > the
> > > > > > tendon will be shorter than the calibrated length because
> > in
> > > > the
> > > > > > calibrated length the strain of the tendon for activity
one
> > > > > > is taken into account. In other words: the muscle is
> > calibrated
> > > >
> > > > for
> > > > > >
> > > > > > activity one. Since we in both cases drive the overall
> > lengh
> > of
> > > >
> > > > the
> > > > > >
> > > > > > muscle to the same length the passive element will be
more
> > > > > > strecthed
> > > > > > in the case where we have the small mass, because the
> > tendon
> > is
> > > >
> > > > not
> > > > > >
> > > > > > loaded so much.
> > > > > >
> > > > > > So you may ask why did we choose to take into account the
> > > > tendon
> > > > > > strain when we calibrate the muscle?
> > > > > > This we did because we wanted to have the optimum
strength
> > > > exactly
> > > > > > at the calibrated length this would not have been the
> > result
> > > > > > otherwise.
> > > > > >
> > > > > > The passive element should work exactly as you describe
in
> > the
> > > > end
> > > > > > of your email, but when the muscle is not loaded very
much
> > the
> > > > > > passive force will not be zero at the calibrated length.
> > > > > > Furthermore
> > > > > > the parameters of your model enhance this phenomean
because
> > > > your
> > > > > > tendon length is extremely long.
> > > > > >
> > > > > >
> > > > > > I hope this answers your question otherwise please do
not
> > > > hesitate
> > > > > > to write again.
> > > > > >
> > > > > > Best regards
> > > > > > AnyBody Support
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > — In anyscript@yahoogroups.com, “Leng-Feng Lee”
> > > > <lengfenglee@>
> > > > > >
> > > > > > wrote:
> > > > > > >
> > > > > > > Hello,
> > > > > > >
> > > > > > > I have a muscle modeled using AnyMuscleModel3E, it is
> > connect
> > > >
> > > > > > through
> > > > > > > viapointmuscle and the distance between the two points
it
> > > > > > connects
> > > > > > is
> > > > > > > 0.5m. I gave the following properties for it:
> > > > > > >
> > > > > > > AnyMuscleModel3E MuscleModelC = {
> > > > > > > F0 = 100;
> > > > > > > Lfbar = 0.5;
> > > > > > > Lt0 = 0.5;
> > > > > > > Epsilonbar = 0.05;
> > > > > > > Gammabar = 30*pi/180; //Pennation Angle
> > > > > > > Fcfast = 0.4; //Fraction of fast twitch
> > > > fibers
> > > > > > > Jt = 3.0; //Elasiticity of tendon
> > > > (serial)
> > > > > > > Jpe = 3.0; //Elasiticity of tendon
> > > > (Parallel)
> > > > > > > K1 = 2;
> > > > > > > K2 = 8;
> > > > > > > PEFactor = 5;
> > > > > > > };
> > > > > > >
> > > > > > > It contract to lift up a segment connected to a joint-
no
> > > > external
> > > > > >
> > > > > > load.
> > > > > > >
> > > > > > > My problem is that the passive force appear throughout
> > the
> > > > > > > simulation—something not desired.
> > > > > > >
> > > > > > > I calibrated the muscle at the initial position, and
try
> > to
> > > > > > change
> > > > > > the
> > > > > > > Lfbar and Lt0 value, but the passive force still exist
> > > > throughout
> > > > > >
> > > > > > the
> > > > > > > simulation.
> > > > > > >
> > > > > > > I would like to have a passive force to act in the
> > following
> > > > way:
> > > > > >
> > > > > > at
> > > > > > > the initial stage, there is no passive force, as the
> > muscle
> > > > > > contract,
> > > > > > > the passive force store up(like compressing a spring)
and
> >
> > act
> > > >
> > > > in
> > > > > > a
> > > > > > > direction opposing the contracting force(?), and as the
> > > > motion
> > > > > > change
> > > > > > > direction, the passive force release(like let go a
> > compressed
> > > >
> > > > > > spring).
> > > > > > >
> > > > > > > Could this be done? What parameter(s) should I
> > change/adjust?
> > > >
> > > > Or
> > > > > > my
> > > > > > > concept about the muscle modeling is wrong? Detail
> > > > explaination
> > > > > > > greately appreciated.
> > > > > > >
> > > > > > > Thanks,
> > > > > > > Leng-Feng
> > > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > >
> > > > >
> > > > > __________________________________________________
> > > > > Do You Yahoo!?
> > > > > Tired of spam? Yahoo! Mail has the best spam protection
> > around
> > > > > http://mail.yahoo.com
> > > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > >
> > > __________________________________________________
> > > Do You Yahoo!?
> > > Tired of spam? Yahoo! Mail has the best spam protection
around
> > > http://mail.yahoo.com
> > >
> >
> >
> >
> >
> >
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
>